DELTA home

The families of flowering plants

L. Watson and M. J. Dallwitz

The first version of these notes was published on the Taxacom Discussion List on 27 April 1999 — see taxacom.markmail.org.

Notes on the APG classifications

The latest edition of this package incorporates the classification of Flowering Plant Families presented by the Angiosperm Phylogeny Group (APG III, 2009), which reflects recent, far-reaching molecular phylogenetic studies. Their ordinal and (formal and informal) supra-ordinal groups are unaccompanied by group descriptions, but inclusion in our package makes them easily accessible (along with earlier schemes) for preparing these by application of Intkey and other DELTA facilities to our compiled family descriptions. The appended natural-language descriptions for the ‘Asterids’ and ‘Rosids’ series of the APG’s 1998 version of the ‘Core Eudicots’ (Ann. Missouri Bot. Gard. 85, 531–553), which were generated using Intkey and Confor, should suffice to exemplify the possibilities. To conserve space, they were restricted to a small suite of the characters showing different tendencies in these two groups.

The SUMMARY option of Intkey generates character-state distributions for any grouping of families, using all the available characters or selected character suites. This takes no more than a minute or two for any grouping that is directly built into the current package, and preparing one for any other series of families requires only that the user first selects their names. Automatic conversion of the data into natural-language descriptions, in RTF or HTML format, may be performed by applying Confor to the Intkey output.

The following tables were obtained using Intkey. Table 1 compares the 1998 APG ‘Rosids’ and ‘Asterids’ groups with the Tenuinucelli and Crassinucelli of this package (cf. Young and Watson 1970, in which we emphasized the homogeneity of the former and the heterogeneity of the latter), while Table 2 contrasts them with Cronquist’s (1966, 1981) subclasses, in terms of family compositions.

Table 1

 

Tenuinucelli

Crassinucelli

Unassigned

Asterids

98

0

2

Rosids

0

149

0

Unassigned *1

2

31

2

Unassigned *2

8

90

6

Table 2

 

Asteridae

Rosidae

Magnoliidae

Hamamel.

Caryophyll.

Dilleniidae

Asterids

46

29

1

1

0

23

Rosids

0

88

1

18

0

42

Unassigned *1

0

10

0

2

15

8

Unassigned *2

0

25

39

7

15

18

*1 APG Core Eudicot but neither Rosid nor Asterid

*2 APG Oddment family, or basal order, or family of uncertain position at the highest group level; or Eudicot but not core Eudicot; or core Eudicot but neither Rosid nor Asterid

Table 1 shows that the APG Asterids were to all intents and purposes our Tenuinucelli, while their Rosids were our Crassinucelli minus 19 families they assigned to ‘basal orders’ and numerous others they had left unclassified. The Asterids also compared quite well with the Sympetalae (Metachlamydeae, Gamopetalae, Sympetalae) of Nineteenth Century systems (e.g. Bentham and Hooker, 1876; Engler and Prantl, 1892), while the Rosids were recognisable as a restricted version of Engler’s Archichlamydeae. By contrast, Table 2 shows how poorly Cronquist’s pseudo-phylogenetic subclasses Asteridae, Rosidae, Dilleniidae, etc., compared with the developing APG groupings, and the two tables clearly demonstrated that ‘Asterids’ and ‘Rosids’ are infelicitous names. We have not bothered to pursue this matter further, other than to confirm that assignments in the formalised 2009 APG classification of families previously unassigned by them merely lends further support to the Tenuinucelli. It is a shame, therefore, that APG continues to use them, instead of acknowledging nomenclaturally the astuteness of the Nineteenth Century taxonomists who detected important groups without the benefits of embryology and molecular biology, and with little or no regard for evolution.

Sample Descriptions

Where not all families of a group are recorded for a character, fractions indicate ‘number of families for which the character state is recorded / the number for which the character is recorded (known) and applicable’. Intrafamilial variation results in the sum of the numerators of all the states of a character exceeding the denominator. These examples incorporate minor editing of the Confor output, to emphasize the more obvious inter-group differences. Ordinal descriptions prepared in the same way would no doubt display less intra-taxon variability than is apparent at higher levels.

Asterids (100 families)

Leaves stipulate (17), or exstipulate (96). Flowers tricyclic (3/75), or tetracyclic (70/75), or pentacyclic (12/75), or polycyclic (4/75). Free hypanthium present (14/86), or absent (81/86). Hypogynous disk extrastaminal (1/11), or intrastaminal (11/11). Perianth with distinct calyx and corolla (95), or sequentially intergrading from sepals to petals (1), or sepaline (13), or petaline (7), or vestigial (2), or absent (3). Corolla polypetalous (43/96, but then often only occasionally so), or partially gamopetalous (1/96), or gamopetalous (72/96); unequal but not bilabiate (23/81), or bilabiate (25/81), or regular (70/81). Androecial members free of the perianth (53/98), or adnate (64/98). Staminodes (when present) external to the fertile stamens (3/20), or in the same series as the fertile stamens (17/20), or internal to the fertile stamens (1/20). Stamens 1–9–‘many’; alternating with the corolla members (50/56), or opposite the corolla members (4/56), or both alternating with and opposite the corolla members (5/56).

Pollen grains 1–4–40 aperturate. Gynoecium 1–4–‘many’ carpelled. Ovules unitegmic (71/82), or bitegmic (14/82); tenuinucellate (73/80), or crassinucellate (13/80), or pseudocrassinucellate (2/80). Outer integument (when present) contributing to the micropyle (3/13), or not contributing to the micropyle (10/13). Endosperm formation cellular (56/74), or nuclear (22/74), or helobial (1/74). Embryo chlorophyllous (11/42), or achlorophyllous (35/42).

Iridoids detected (52/85), or not detected (33/85). Ellagic acid present (9/70), or absent (65/70).

Rosids (149 families)

Leaves stipulate (98/148), or exstipulate (92/148). Flowers tricyclic (6/69), or tetracyclic (42/69), or pentacyclic (40/69), or polycyclic (15/69). Free hypanthium present (54/121), or absent (84/121). Hypogynous disk extrastaminal (17/38), or intrastaminal (26/38). Perianth with distinct calyx and corolla (120), or sequentially intergrading from sepals to petals (3), or sepaline (53), or petaline (5), or of ‘tepals’ (1), or vestigial (14), or absent (16). Corolla (when present) polypetalous (115/119), or partially gamopetalous (3/119), or gamopetalous (36/119, but then often only slightly or rarely so); unequal but not bilabiate (19/83), or regular (72/83). Androecial members free of the perianth (132/140), or adnate (27/140). Staminodes (when present) external to the fertile stamens (17/28), or in the same series as the fertile stamens (8/28), or internal to the fertile stamens (9/28). Stamens 1–21–‘many’; alternating with the corolla members (14/39), or opposite the corolla members (9/39), or both alternating with and opposite the corolla members (24/39).

Pollen grains 2–5–‘many’ aperturate. Gynoecium 1–5–‘many’ carpelled. Ovules unitegmic (13/121), or bitegmic (114/121); tenuinucellate (14/112), or crassinucellate (103/112). Outer integument (when present) contributing to the micropyle (68/85), or not contributing to the micropyle (36/85). Endosperm formation cellular (6/98), or nuclear (94/98), or helobial (3/98). Embryo chlorophyllous (40/61), or achlorophyllous (31/61).

Iridoids detected (4/83), or not detected (80/83). Ellagic acid present (40/94), or absent (70/94).


Contents